Baby-boomers can expect better imaging of age-related changes

June 1, 2007

If ever a generation has been obsessed with aging, it's the baby-boom generation. Thirty years' worth of fads in exercise, diets, fashion, and personal grooming have all been directed at slowing the aging process or hiding its effects. Well, boomers can run (or jog, or bike, or yoga), but they won't be able to hide from what radiologists may soon be able to tell about the age of their insides.

If ever a generation has been obsessed with aging, it's the baby-boom generation. Thirty years' worth of fads in exercise, diets, fashion, and personal grooming have all been directed at slowing the aging process or hiding its effects. Well, boomers can run (or jog, or bike, or yoga), but they won't be able to hide from what radiologists may soon be able to tell about the age of their insides.

As the elderly population in the U.S. grows, radiologists are increasingly being called on to perform imaging studies of age-related acute and chronic diseases. An atlas that will document a baseline benchmark of normal age-related changes in the body and how they relate to or influence and support age-related diseases is being compiled by researchers at the University of Pennsylvania.

Spearheading the project is Dr. Abass Alavi, emeritus chair of the department of nuclear medicine at Penn and a pioneer in FDG-PET imaging. He and his colleagues have undertaken a massive project to quantitatively document the normal aging process with PET, CT, and MR imaging.

So far, Penn researchers have published 11 studies in Seminars in Nuclear Medicine. More studies with additional detail are slated to appear this month and in the future. The effort involves retrospectively reviewing images from a representative cross-section of the population, including pediatrics, all considered healthy. Researchers recorded organ volumes and metabolic activities in the brain, head and neck, thorax, heart and blood vessels, breast, abdomen, pelvis, bone marrow, and skin, fat, and muscle. With this knowledge, nuclear medicine physicians and radiologists will be better able to differentiate early pathology from changes of normal aging.

"Despite the fact that everyone undergoes aging, there is very little in the literature regarding quantitative assessment of normal age-related changes in structure and function of the different organ systems of the body through use of imaging," said Dr. Drew A. Torigian, an assistant professor of radiology at Penn and senior author on several of the papers published in this effort. "In order to know whether a structure is normal or not, you need to first know what's considered normal."

The thymus gland, for instance, shrinks with age. In a 50-year-old adult, radiologists generally would not see the thymus on a PET or CT scan. But in a 10-year-old child, they'd find a large thymus with some metabolic activity. Both are normal, Torigian said.

Imaging atlases of normal structures already exist, but this current effort differs in its organized assessment of changes that occur in normal aging and in the use of precise quantitative measurements, Alavi said. Radiologists have been slow to adapt age-related knowledge into their practices, and they have not fully embraced quantitative analysis.

"The two things we want to communicate with this atlas are the age-related changes associated with imaging and the need for radiologists to use quantitative measurements to more accurately depict pathology and its concomitant treatment or interventions," he said. "We have the tools and they're more advanced than ever, so we should use them."

The Penn team also introduced a new parameter that incorporates both structural and functional information into a single unit called the metabolic volumetric product, or metabolic burden. To arrive at the metabolic burden, the volume of the structure is multiplied by the standardized uptake value of FDG. The resulting number incorporates the organ's volume and metabolic activity, which could be an important global disease activity parameter for gauging therapeutic effects on cancer and other disorders (Semin Nucl Med 2007;37[3]:223-239).

"We've made some plots of the metabolic burden in several organs, such as the liver and spleen, to show that it might be a useful parameter in the future for any disease pathology," Torigian said.


Alavi and colleagues at Penn pioneered brain imaging in the 1980s, using MRI and PET to show the aging brain. They were some of the first researchers to show that functional changes preceded structural changes. Their latest data validate earlier findings (Semin Nucl Med 2007;37[2]:69-87). They forged the use of volumetric segmentation, separating gray/white matter and cerebrospinal fluid to quantify specific metabolic activity within those areas. One of the big problems with functional imaging is its low spatial resolution, which results in an underestimation of FDG activity, a phenomenon called partial-volume effect. Researchers correct for this by combining PET with segmented regions of interest from high-spatial-resolution structural imaging techniques.

Many diseases of the brain involve either gray matter or white matter, and having precise metabolic measurements of specific regions ultimately leads to better patient management, Alavi said. Precise metabolic measurements become especially important when gauging the effects of a new drug or therapy. Without correcting for partial-volume effect, investigators may not see the true outcome of a drug. They will underestimate its effect, and research could be prematurely halted.

A similar quantitative approach can provide an accurate calculation of thoracic muscle metabolic burden, which tells clinicians which muscles are contracting and which are not, an indication of the lung's functional ability (Figure 1). These data can help determine the effects of drugs as the rush to develop effective treatment for airway diseases intensifies, he said.

The group documented the lung's normal aging process as a benchmark for future investigations (Semin Nucl Med 2007;37[2]:103-119). Led by David S. Well, MSE, from the radiology department, researchers found that lung attenuation (measured in Hounsfield units) and SUV decreased significantly with age, while lung volume remained stable (Figure 2). They also found that lung metabolic burden (volume x SUV) remained stable with age.

There was a correlation between body mass index and other processes. Lung attenuation, for example, increased with increasing BMI, which subsequently correlated with an increase in lung SUV (Figure 3). But heavier people also tended to have reduced lung volume, which suggests restriction of the rib cage or lungs by excess fat deposition. With an increase in metabolism (SUV) and a decrease in volume, one would expect to find a relatively stable metabolic burden as BMI increased, which is exactly what Well and colleagues found.


The idea of combining structure and function to arrive at a quantitative global measurement of disease activity can be generalized to all organs, according to the researchers.

"The concept of metabolic burden will enhance the overall effects of disease management rather than what we do by 'punch biopsy,' which is taking only a small region of interest of an organ and talking about an entire organ," Alavi said.

The use of tissue segmentation was especially important to accurately calculate the metabolic activity of bone marrow. Although FDG is not a bone marrow-specific tracer, it is useful for examination of functioning red marrow as well as detection of bone marrow involvement by both benign and malignant disorders. One of the group's studies (Semin Nucl Med 2007;37[3]:185-194) is the first to use FDG-PET to describe physiological distribution of the marrow and to assess the changes in red marrow that occur with normal aging over a wide age range, according to the authors. The results are preliminary and researchers intend to conduct further prospective studies (see figure, page 33, top).

Drs. Judy S. Blebea and Mohamed Houseni, co-lead authors from the radiology department, and colleagues examined five subjects using segmentation on MRI to correct for PET's partial-volume effect.

  • The mean volume of the lumbar vertebral body was 15.6 cm3; average maximum bone marrow SUV was 1.5; and metabolic burden was 23.4 (SUV-cm3).

  • The mean volume of the yellow marrow in the lumbar vertebral body was 7.7 cm3; average yellow marrow SUV was 0.38; and metabolic burden was 2.9 (SUV-cm3).

  • The mean volume of the red marrow was 7.9 cm3; red marrow SUV was 2.6; and metabolic burden was 20.5 (SUV-cm3).

Researchers also undertook a more comprehensive qualitative look at bone marrow activity and distribution in 112 patients. They divided patients into eight age categories with equal gender distribution in each: 0 to 15 years and then in increments of 10 years up to 85. They found no significant difference in SUV in the red marrow activity and distribution between men and women in the same age group. While the upper arm and leg bones showed no significant difference with age, FDG uptake in the extremities declined significantly with advancing age. Bone marrow SUV in the manubrium and the 12th thoracic and fifth lumbar vertebrae demonstrated a weak correlation with advancing age.

"We believe that the combination of these approaches will ultimately enhance the role of medical imaging in the management of primary and secondary marrow disorders," Blebea and Houseni said.

While qualitative changes in breast composition are well established, quantitative data are scarce. Drs. Richard G. Abramson and Ayse Mavi of the radiology department and colleagues used the same segmentation techniques for accurate quantitative anatomic and functional breast data (Figure 4). They retrospectively examined breast images from digital mammography, CT, MRI, and FDG-PET. Not surprisingly, they found that the volume of glandular tissue decreases with age. But this is the first study, according to the authors, that shows a correlation between age and FDG uptake in the breast tissue, which also decreased with advancing age (Semin Nucl Med 2007;37[3]:146-153).

Because certain breast composition patterns carry a higher risk of developing breast cancer, the authors suggest that normal age-related variations in breast density could aid in developing and refining risk prediction models. The authors also suggest that the metabolic burden could personalize breast cancer screening. Women with more glandular tissue or abnormally high baseline metabolic activity may receive earlier or more frequent mammographic screening, or they may be triaged to screening with ultrasound or MRI.


The Penn team also coined the term atheroburden, a product, in this case, of the aortic wall volume and its FDG uptake (Nucl Med Biol 2006;33[8]:1037-1043). In this retrospective study, Dr. Gonca Bural of the radiology department and colleagues showed that the volume of the aortic wall increased with age as it thickened with plaque. Metabolic activity also increased, as a function of the inflammatory cells in plaque, but at a slower rate than the volume. Overall, the atheroburden increased with age. Alavi suggested that clinicians could use the atheroburden to gauge the effects of statins and other drugs that serve to control or reduce plaque.

Based on this research, cardiovascular specialists at Penn have begun to talk about changing their protocols to include the atheroburden, which may help determine whether statins are working. This is exactly what Alavi wants to see done with the data. Neither he nor any member of his team stands to profit from this endeavor. The research is meant to stand as a barometer for disease activity and to be used by researchers and clinicians worldwide.

There are implications for every disease, including those of the reproductive organs. Led by Well and Dr. Hua Yang, researchers analyzed age-related changes in the uterus, ovary, testicle, and prostate gland. Their results, while generally consistent with previous findings, added precise quantitative data on the effect of aging on these organs (Semin Nucl Med 2007;37:173-184).

They found that the mean volume of the uterus increased for women from their 20s to their 40s, then began a progressive decline well into the eighth decade of life-a 60% reduction from peak volume. Ovary volumes remained relatively stable until late in the fifth decade of life and were not identifiable on MRI in subjects beyond the sixth decade, a statistically significant decrease.

As more women choose to get pregnant later in life, it's important to understand the effects of aging on their reproductive function. Such morphologic and functional benchmarks can help physicians appropriately advise and treat this growing population.

The volume of the prostate and its central gland increased significantly from the second decade of life to late in the eighth decade (Figure 5). Testicular volume rapidly increased with age from birth to 25 years and then declined significantly. Maximum SUV increased as well up to age 35, plateaued until age 65, and then de-creased slightly.

In a retrospective study of both adult and pediatric abdominal organs, Jeffrey M. Meier, MSc, and colleagues found that volumes of adult liver, spleen, pancreas, and kidneys did not change significantly with age. However, their attenuation decreased significantly (Semin Nucl Med 2007;37:154-172).

In contrast to previous studies, researchers found a significant increase in liver metabolism with age, although the overall metabolic burden did not change significantly. They suggest the increased hepatic FDG uptake could reflect inflammatory changes, which might also explain the age-associated reduction in liver attenuation.

Alavi suggested that functional data on muscle contraction in the abdomen can predict diseases such as diabetes. Those data in the small bowel and colon can potentially help with early detection of diseases such as irritable bowel syndrome, Crohn's, and colon cancer.

Many of these findings come from ongoing research activities that are funded by the National Institutes of Health. Researchers have already asked the NIH to fund prospective studies. Alavi says he is confident the research will be funded as the "interest level is high in aging and imaging studies."

While the aggregate of these studies from the University of Pennsylvania lays the groundwork for creating a database of normative structural and functional values associated with aging, an equally important aspect of the research involves efforts to take quantitative techniques that are reasonably accurate and simplify them enough for use in daily practice. The importance of PET and PET/CT in routine clinical use to diagnose, stage, and monitor treatment has grown significantly. As the focus of medicine shifts from the structural to the molecular and from the collective to the personal, the need for sophisticated quantitative techniques that can document the smallest changes will become even greater.

Mr. Kaiser is news editor of Diagnostic Imaging.